A. Casimiro and P. Ver ssimo. Timing failure detection with a Timely Computing Base. In Proceedings of the European Research Seminar on Advances in Distributed Systems, Madeira Island, Portugal, Apr. 1999.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....by a real time unreliable channel (messages can be lost or duplicated) Messages delays have a distribution similar to the one depicted in Figure 2. Thus, if deadlines are reasonably selected, most messages will meet those deadlines. Nevertheless, the channel may exhibit occasional timing failures [6] if the chosen deadline is smaller than the absolute worst case delay (TD max , in the figure) For sake of clarity, we assume that processing steps take no time, only message transfer delays consume time. In this paper we are concerned with the class of applications with timeliness requirements ....

A. Casimiro and P. Ver'issimo. Timing Failure Detection with a Timely Computing Base. 3rd European Research Seminar on Advances in Distributed Systems (ERSADS'99), Madeira Island, Portugal, April 23-28, 1999.

No context found.

A. Casimiro and P. Ver ssimo. Timing failure detection with a Timely Computing Base. In Proceedings of the European Research Seminar on Advances in Distributed Systems, Madeira Island, Portugal, Apr. 1999.

....actions requires a protocol to be executed by all TFD modules on top of the control channel. Recall that we are talking about actions in the payload system. For lack of space, in this paper we only provide the intuition behind this protocol. However, the protocol is available in a technical report[5]. The protocol executes in rounds, during which each TFD instance broadcasts all relevant information for the detection of timing failures. Distributed common knowledge of failure occurrences is achieved by disseminating information about locally detected failures. When an application sends a ....

....message, the TFD service uses the distributed duration measurement service to measure its delivery delay. Therefore, TFD instances in all destination sites will collect the measured value. The error associated to this measurement yields the value of ) specified in property TCB5 ( [5] for a proof) The TCB of the sender site logs the timed action specification of the sent message for later use. Each send request is tagged with a unique identifier, known to the TFD service and to the application. The information concerning each sent message (identifier and bound specification) ....

A. Casimiro and P. Verssimo. Timing failure detection with a timely computing base. In Third European Research Seminar on Advances in Distributed Systems, Madeira Island, Portugal, May 1999. Available as Tech. Report, Department of Informatics, University of Lisboa, DI/FCUL TR-99-8.

....absolute timestamping Provides globally meaningful timestamps. Figure 2.5: TTCB Services The TTCB services are summarized in Figure 2.5. The implementation of the security related services is presented in [27, 26] The design and implementation of the time related services can be found in [70, 19, 20]. The following subsections describe the TTCB services and their APIs. Figure 2.6 shows the meaning of some common API parameters. Parameter Description eid an entity identification before the TTCB elist a list of entities identified by their eids tag an unique id for an execution of a service ....

....or (M ndlv.number M.higherseqVector[M ndlv.sender] then 12 R deliver(M ndlv.data) remove M ndlv from notDelivered; 13 if (higherseqVector[M.sender] M.number) then 14 R deliver(M.data) 16 put M in notDelivered; higherseqVector[M.sender] M. number; 17 round r : round r 1; line [20]) The message has an header with the sender identifier, a sequence number and the table higherseqVector. This table has an entry for every local TTCB that contains, for every other local TTCB, the highest sequence number of a message received from that local TTCB. The receive routine starts by ....

A. Casimiro and P. Verssimo. Timing failure detection with a Timely Computing Base. In Proceedings of the European Research Seminar on Advances in Distributed Systems, April 1999.

....actions requires a protocol to be executed by all TFD modules on top of the control channel. Recall that we are talking about actions in the payload system. For lack of space, in this paper we only provide the intuition behind this protocol. However, the protocol is available in a technical report[5]. The protocol executes in rounds, during which each TFD instance broadcasts all relevant information for the detection of timing failures. Distributed common knowledge of failure occurrences is achieved by disseminating information about locally detected failures. When an application sends a ....

....sends a message, the TFD service uses the distributed duration measurement service to measure its delivery delay. Therefore, TFD instances in all destination sites will collect the measured value. The error associated to this measurement yields the value of T TFDmin specified in property TCB5 (see [5] for a proof) The TCB of the sender site logs the timed action specification of the sent message for later use. Each send request is tagged with a unique identifier, known to the TFD service and to the application. The information concerning each sent message (identifier and bound specification) ....

A. Casimiro and P. Verssimo. Timing failure detection with a timely computing base. In Third European Research Seminar on Advances in Distributed Systems, Madeira Island, Portugal, May 1999. Available as Tech. Report, Department of Informatics, University of Lisboa, DI/FCUL TR-99-8.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC